Quick Summary. Google’s street view is simply a representation of reality on a specific day, and they have not highlighted any aspect of the dataset, and furthermore the dataset is comprehensive. Given the mapping between reality and the dataset that is inherent in something like Street View, one’s privacy on the day your photo was taken and one’s privacy in the dataset are commensurate, because your relative anonymity is the same in each. Arguments pointing out that certain people and websites can highlight compromising pictures are missing the point, and are like blaming camera manufacturers for the actions of paparazzi. If a company decides to single out a certain picture on somebody on Street View on your website, that company is the party violating privacy, not Google. Google isÂ producing an unbiased representation ofÂ reality; just as in physical reality, it is the choices and actions of others who decide whether or not privacy is violated.
Recently, Google has been driving around various metropolitan areas (including Boston) in a fleet of funky-looking cars adorned with eight cameras mounting on their roofs (see below)Â profligately photographing everything within view of the street every few feet, and linking the resulting panoramic shots to their respective locations in Google Maps. Their eventual goal is to have virtually every building on every street in every major city photographed, such that you can click on a street and see a picture of the surroundings from that location. You’d have to be Mr. and Mrs. Boring to not think that’s cool.
Right now, the resolution is sufficient to find that bar from which you stumbled home one night but whose name eludes, or to get a decent idea about whether or not the Lake View Retirement Home really has one. As it grows more complete, it will be a profoundly powerful dataset, and will doubtless result in all manner of unforeseen applications. This will be especially true if Google actually uses higher resolution pictures. Do you want to see when a favorite business is open, but they don’t have a website? You could, in theory, check out the hours posted on the front of their store with sufficiently high resolution imagery. If you’re wondering about the legal parking hours on the streets near a restaurant you’re planning to visit, you could read the parking signs across town from your computer.
Unfortunately, reactionary privacy concerns have plagued the service since its inception, and if the service survives at all, it’s likely that it will be limited to low resolution pictures. Some of the criticism has predictably come from people who have been photographed doing things they shouldn’t, but much of the ire has come from people who simply think that having a picture of them taken while they were in public shouldn’t be allowed online. And I have to admit, I took pause when I found our own car parked in our usual spot:
However, upon further reflection, I realized that it is unreasonable to object to this as a privacy violation, for reasons that are especially clear in this particular case. Quite literally, there is going to be a nearly one-to-one correlation between the Google Street View dataset and the real world. Thus, while your image might be available to everybody, so are millions of other images. One might expect that at any given moment, the proportion of people interested in the Google picture of the specific place you were the day the google car spotted you is very roughly the same as those interested in that specific spot in real life at any given moment. Thus, for exactly the same reason you only saw a few people on the street with you at the moment the picture was taken, it’s likely only a few people are interested, at any one time, in that picture.
As such, your relative obscurity, and therefore privacy, in Google Street View exists for the same reason it exists in real life: the whole world is not going to suddenly look at the picture of you for the same reason the whole world didn’t spontaneously decide to take their vacation to Prescott Avenue the day my car happened to be parked there. People focus on the fact that, technically, their picture is accessible to the billion or so users of the Internet. But that’s not really relevant. Technically, most of those people could also afford to travel to your old highschool’s library to look at the picture of you in the archival yearbook. What is relevant is the probability of any of them doing so. Thus, while I understand that it’s psychologically unsettling to have a picture of yourself on the internet taking out the garbage shirtless, probability dictates that you aren’t likely to be spotted by too many more people than were able to see you that day, anyway. (Unless, of course, you decide to be a schmuck and sue Google, thus bringing highly ironic media attention to yourself.)
Furthermore, this is nothing new. There has long been the possibility of finding yourself in the background of a snapshot that somebody has placed online, and people seem to readily accept this, perhaps because of their familiarity with the concept of photography. Perhaps the novelty of the Street View application is to blame for people’s skepticism. Or, perhaps it’s simply that a big, mean, corporation is behind it that leads them to think of this as different. But for whatever reason, publicly displaying pictures of a public scene taken from a public road has caused a tremendous hue, despite this being nothing new. My hope is that Google will not bow to the pressure, because I’m sure that people will eventually aclimate to this new “invasion” of our privacy, just as their predecessors in worry did over a hundred years ago when they realized the potential horror of being spotted in the background of a picture in the New York Times.
All of this leads me to the following proposal for what should constitute an online violation of privacy: An expectation of online privacy should be commensurate with your expectation of real-life privacy at the moment your personal information or likeness was captured. This means that the “primacy” of your likeness in the data set in which you find yourself (i.e. it’s significance relative to the total of the dataset) should be commensurate with your visibility in the location at which you were captured. For example, if you’re walking down the street of your small town, perhaps it’s fair that your picture is to be found among a data set that contains virtually every street in America, as will eventually be the case with Google Street View. On the other hand, finding the same picture blown up on the main page of www.cnn.com would almost certainly be a breach of your expectation of privacy, as your prominence online would then be completely out of proportion with your prominence on the day the photo was taken.
One likely objection to my argument is that having your picture taken by Google could potentialy lead to prominent publication outside the scope of the Street View dataset. Wouldn’t that then be a violation of privacy, one for which Google is ultimately liable? True, perhaps it would be an invasion of privacy, but that doesn’t mean Google should be held accountable. Again, the same principle of correlation between the dataset and real world applies. If you gather a subset of the Street View dataset (i.e. a picture of a person) and publish it prominently, that is akin to taking a photograph of somebody and publishing it. This may seem an odd analogy to make, as the physics may not be at all the same, but the informational aspect of the actions are identical: in both cases you are focusing on and promoting one element of a large dataset. In the first case the “dataset” was real life, and in the second it was the Street View database, but in both cases the subject was anonymous until action was taken to make it otherwise. Thus, the person who extracted the Street View image and made the effort to publish it prominently should be the person liable for the privacy violation, not Google. Holding Google to blame in the first example would be like suing the manufacturer of the camera used in the second.
We have to stop treating publication on the internet as a black-or-white phenomenon when it comes to privacy. It is rapidly gaining the ubiquity of physical presence in terms of our communications, and our expectations for online privacy will have to adopt some shades of grey if we are to avoid a litigious free-for-all. Just as the level of one’s privacy varies by location in the real world, we will have to begin to see “locations” on the internet as falling along a similar continuum. That is, just as there can be relative degrees of anonymity in physical space, we must recognize that the same concept can exist in virtual spaces.